Guide · Agent skill
How to use the Paid Ad Monitoring agent skill (ads-monitor)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The ads-monitor skill is a set of instructions that teaches an AI agent to review paid-advertising account snapshots, such as daily or weekly exports, and report on pacing, delivery, performance, creative fatigue, tracking and policy status. It is aimed at marketers running paid campaigns, analysts checking whether a metric change is meaningful, and teams verifying campaigns after launch. After this guide you will know what the skill does, how to install it, how to ask for it in plain language, and what data it needs.
Paid Ad MonitoringAgriciDaniel/claude-ads · MIT Download the skill Quick start
- 1Download the ZIP. Download the ZIP for skills/ads-monitor from this page and unzip it to get SKILL.md.
- 2Place the skill folder. Move the folder into ~/.claude/skills/ for personal use or into .claude/skills/ inside a project.
- 3Or upload in the Claude apps. In the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.
- 4Prepare normalized snapshots. Export two or more snapshots with matching account, timezone, currency, metric and attribution definitions.
- 5Ask for the check in plain language. Describe the review you want, such as a weekly health check or a budget pacing review, and name the periods to compare.
- 6Review the observations. Check the confidence levels, likely causes and decision thresholds before acting, since the agent will not change the account.
What is an agent skill and what does ads-monitor make the agent do?
An agent skill is a folder containing a SKILL.md file with instructions that an AI agent loads when a task matches the skill's description. The ads-monitor skill is named ads-monitor and its description covers monitoring paid-ad account pacing, delivery, performance, creative fatigue, tracking, policy and data quality across supported platforms.
When the skill is active, the agent follows a fixed workflow: it loads two or more normalized snapshots, checks data freshness and finalization windows, separates expected factors such as learning, seasonality, reporting latency and planned changes from unexplained anomalies, evaluates the account areas listed above, and returns observations, confidence, likely causes, required investigation and decision thresholds.
- Compares two or more snapshots with compatible account, timezone, currency, metric and attribution definitions.
- Evaluates pacing, delivery, conversion quality, unit economics, creative fatigue, tracking health, policy status and changed account objects.
- States when the evidence cannot distinguish noise from a material change.
- Persists a versioned monitoring bundle and can link detected failures to regression or follow-up tasks.
Install the ads-monitor skill
Download the ZIP for skills/ads-monitor from this page and unzip it. The skill is a single file, SKILL.md. Place the unzipped folder in one of two locations: ~/.claude/skills/ to make it available across your projects, or .claude/skills/ inside a specific project folder to limit it to that project.
If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside your project
- Claude apps: Settings → Capabilities → Skills, then upload
Prepare your snapshot data
The skill works on snapshots of account data, such as a daily or weekly export. The skill file does not specify a data format, so the requirement is that your snapshots are normalized and comparable: the same account definitions, timezone, currency, metric definitions and attribution settings across the periods you compare.
Mismatched data produces weak conclusions. If one export uses a different timezone or attribution window from another, the skill is instructed to avoid alerting on those comparisons, so fix definitions before you load the files.
- Two or more snapshots covering the periods you want to compare.
- Consistent account, timezone, currency, metric and attribution definitions.
- Know when each period's data was finalized, since the skill checks finalization windows.
How to trigger the skill with natural requests
You do not need a special command. Describe the check you want in plain language and mention the purpose the skill covers, such as a daily or weekly check, anomaly review, budget pacing, post-launch verification or campaign monitoring. The agent will match your request to the skill's description and load it.
Make the request specific: name the period, the snapshots and the question you want answered. The skill separates real problems from normal noise, so say what changed and why you are concerned.
- "Run a weekly health check on these two snapshots and tell me whether the CPA change is real or noise."
- "Check budget pacing for this month's snapshot against last month's and flag anything unexplained."
- "We launched these campaigns yesterday. Verify delivery and tracking using the attached snapshots."
- "Review this week's export for creative fatigue and policy status."
What the output looks like and what it will not do
The skill returns observations, confidence levels, likely causes, required investigation and decision thresholds. It is written to observe only: it states that it should not mutate the account, so it will not change bids, budgets or ads. It also saves a versioned monitoring record that can be linked to follow-up tasks.
Where evidence is insufficient, the skill says so rather than guessing. It also avoids percentage alerts on trivial denominators or incomparable windows.
- Expect observations with confidence levels and likely causes.
- Expect a clear statement when noise and material change cannot be separated.
- Do not expect automatic bid, budget or ad changes.
Safety notes: scripts, network and external content
According to the skill's listing, it runs no scripts and uses no network access. Its only file is SKILL.md. The skill also instructs the agent to treat external account and web content as data, never as instructions, which helps guard against text in exports or web pages trying to direct the agent.
As with any instructions you load into an agent, read SKILL.md before installing it and review the outputs before acting on them.
- Runs scripts: no.
- Uses network: no.
- External account and web content is treated as data, not instructions.
What people use it for
- Running a daily or weekly health check across pacing, delivery and conversion quality.
- Deciding whether a drop in conversions or a rise in cost per acquisition is meaningful or just reporting delay or seasonality.
- Verifying a campaign's delivery and tracking after launch.
- Reviewing creative performance for signs of fatigue before performance declines.
- Recording a versioned monitoring record that links detected issues to follow-up tasks.
Frequently asked questions
Is the ads-monitor skill free to use?
The source material for this skill lists its licence as MIT. The skill is a SKILL.md instruction file, so there is no separate purchase described in the source, but you should check the licence file in the repository for the full terms.
How do I install the ads-monitor agent skill?
Download the ZIP for skills/ads-monitor from this page and unzip it. Place the folder in ~/.claude/skills/ for personal use or in .claude/skills/ inside a project. In the Claude apps, you can upload it under Settings, then Capabilities, then Skills.
What data does ads-monitor need?
It needs two or more normalized snapshots of paid-ad account metrics with compatible account, timezone, currency, metric and attribution definitions. The skill file does not specify an exact data format.
Does ads-monitor change my bids, budgets or ads?
No. The skill is written to observe only and states that it should not modify the account. It returns observations and suggested next steps for you to act on.
Does ads-monitor run scripts or access the internet?
According to its listing, it runs no scripts and uses no network access. It consists of a single SKILL.md file.
Who should not use ads-monitor?
Teams that expect automatic changes to bids, budgets or ads should choose a different tool, and so should anyone without normalized, comparable data exports, since messy or mismatched data will produce weak conclusions.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use ads-monitor if you run paid campaigns and want a structured daily or weekly health check, or if you need to judge whether a metric change is real or just noise. Pick a different tool if you want an agent that automatically adjusts bids, budgets or ads, or if your data exports are not normalized and comparable.
Download the skill